tyre

Maybe you would have. Maybe you wouldn't have. But they have to design these things for the 90% case. People using things that are far outside what's being taught—in the age of AI—are maybe more likely to be cheating than everyone going above and beyond.

Students are human. Many have been taught to focus on one end goal: a grade. They see a mythical program that knows 6000x more than is taught in this 200 level course; of course they want to use it!

I mean people said this all the time, but wouldn’t pay for it (as you said), couldn’t recruit for it, and definitely couldn’t retain them.

But so many teams said they wanted to Raise the Bar to infinity and hire a World Class Team.

They built this brand for hardware, but, personally, I’ve never had this opinion of their software. I mean, being Unix based is better than Windows, but they’ve always been cooking weird UI refreshes or buggy software.

It’s crazy because they have (or had) something like $100bn in cash. Spending 0.5% of that per year on software (development, bug-fixing, performance, and QA) would help tremendously. They’d be insurmountable.

I know “they don’t have to”, but they don’t have to do a lot of things that they do on the hardware side. They choose to.

iirc these aren’t native menus. Firefox builds their own. I found this out when scratching my head why I couldn’t find “Look Up” in the right click menu in Firefox.

I had a fun one where clicks stopped registering sometimes and right clicks didn’t register sometimes and my mouse disappeared randomly.

I thought my machine was broken. Turns out this is a known bug with window managers (????????)

The sporadic nature of it had me question my own sanity.

Because a lot of people employed as engineers can’t build classifiers. They’re able to glue libraries and tools together, build UIs, and write APIs, but don’t have the curiosity or creative problem solving to learn and master a new domain. Even when “master” is scoped to something like this.

On top, most EMs wouldn’t take a risk on an exploration of something “unknown” (to them) and couldn’t get buy-in from a PM.

I say this as an EM. Interview hundreds of people and, while, yes, some people don’t interview well, you might be shocked at the level of creative thinking. Even when “creative” is narrowly scoped to “this is a solved problem in a related domain.

What are the costs compared to an ML model?

You must not have invalidated your cache since a third thing.

> it has no relationship whatsoever with "the origins of life".

followed by

> it can provide information about how the early stages of the hybridization happened

so…

the origins of major parts of life?

Last I’d checked, Brockman is the largest single Trump Super PAC donor in 2026. It’s a midterm, though, and because of the existing focuses (Iran, Fed, etc) there isn’t much that needs bribing at the moment. Pardons cost less than Brockman’s given, no new taxes or reform are likely, and AI is safe.

The largest structural danger for the present economy is interest rates and AI investment (the primary driver of index returns) are (1) tightly linked (2) moving in the wrong directions.

We’ll see whether Anthropic goes through with an IPO and what happens. If the numbers look wrong (even if they’re good, but not good enough), everything could come tumbling down.

Check out asianometry on YouTube for deep dives into many parts of semiconductor manufacturing, including the lithography processes ASML uses.

The difference between Anthropic and a cluster bomb supplier is that cluster bomb manufacturing in the United States requires permits. The government is only going to give them to companies that are working with the United States and its allies, which means that company has zero leverage to make requirements (e.g. "don't use them on humans.")

Separately, as a note to your made up example:

Cluster bombs are uniquely efficient against humans, especially in trenches. The US is not a signatory to the international treaty banning their use. It does not like being limited in how it conducts war. I strongly doubt that it would agree to terms from a private enterprise to similar effect.

The DSA branch has almost zero political power. Yes, Mamdani was elected. No, it was not because of DSA.

DSA has very limited understanding of the politics of power. Agree or disagree with their policy and advocacy, but they lack a unified, coherent policy platform. They take stances from (imo) good places and suggest proposals that are politically untenable (to say the least) and realistically ridiculous.

Mamdani is a unique politician. He is generationally charismatic, articulates the pragmatism behind solutions, and taps into the everyday struggles of citizens. When he talks about ideologically, it's soaring rhetoric about who we can be. His critiques, however, are often strongly tethered to tangible concerns (rent, food, education) vs. the DSA-type far left.

Source: worked in democratic politics, on a presidential campaign, with a wide range of people on the left (and "left") including DSA die hards.

why do they have to decide?

Seems like they overloaded whatever they use for telemetry to do feature flags.

It’s not a crazy conspiracy. They messed up, it’s fine.

iirc the NYC subway is only about 40% subsidized by fares. I’m happy to pay more in taxis for it to exist and subsidize lower income residents. There are huge economic benefits to the city that you’re not factoring in.

Not everything has to be directly money making.

If anything, the Bay Area needs significantly more public transit. BART in particular is mostly useful between cities, in my experience. You can’t go from the mission to cow hollow or dog patch to north beach, let alone anywhere to the ocean.

Also, subways don’t have much (if any) marginal cost per passenger. The issue isn’t cost but ridership. Ridership will increase with utility, which requires political will, which does not exist.

SF could be so much more.

Oh yes, I’m 100% with you. I wish they’d keep it. It’s better for users but perhaps untenable for the business.

WoW Forever will give you a chance for redemption.

Yes, huge amounts of your medical information is for sale. In 2024, Change Healthcare (CHC) was hacked and held ransom. The hackers were in the system for over a week before everything was pulled offline.

CHC is the largest claims clearinghouse in the US; about 100m people's insurance claims go through there each year.

The hackers asked for a ransom ($35m iirc) to delete the data, which United Healthcare (who owns them, because of course they do) paid. But it seems that the collective negotiating on behalf of the actual hackers rug pulled, so the actual hackers didn't get paid.

This is more than simply medical records. It includes who is active duty military and their family. If you can map where soldiers are, you know roughly the size of different military bases. If you know which types of capabilities are operated out of which bases, you can get a pretty good look of how the US is allocating personnel.

It was crazy working on recovery from this at the time. It should have been front page news, but wasn't.

I doubt most users would look at them if they were available. More likely they don’t want to stream distillation material.

I appreciate that you shared it. I didn’t see it at the time.

Wow this is super cool. I wonder when LLMs get cheap enough that NPCs will be able to have full conversations with players.

Not sure they’re the best option for raiding, but as a high-level orchestrator for choosing content, that sounds pretty great.

Spain’s youth unemployment is the highest in Europe. I hope it’s improving, but economically Spain has not recovered post-2008 in line with others in the EU.

IMO it’s well-suited to the general public!

What is the average person’s threat model? Re-using passwords.

What is their most common annoyance? They forget their passwords.

Built-in password managers are great for this. But they’re not universal (e.g. I believe using Apple Passwords on Mac doesn’t carry over to Android.)

Passkeys are amazing. They’re basically magic to the average person. No typing passwords; no remembering; simply vibing, securely.

I don’t think people are regularly logging in to a huge number of accounts on other devices. Maybe they are! I doubt it.

One issue for developers is that the most powerful models refuse to do comprehensive reviews. You can’t ask Fable 5.1 to find every exploit in your codebase, because that’s indistinguishable from what a bad actor would do.

This has always been the case, forever. You have to pay for a product or service. How you do so can be with cash or your data/body/vote/eyeballs/indirect discretionary purchases.

The amount of work that can be done funded by foundations and free work is nowhere close to what people want.

Being a small player and throwing away any hope of expanding your user base by building for people who will complain no matter what you do is, well, something!

> yet for inexplicable reasons

Come on, man, I’m going to choose you’re just Posting saying shit like this.

Because it’s useful. Because almost nobody has the hardware to run a local model with nearly reasonable performance (in intelligence + speed + RAM usage).

Not everything is perfect. At least they’re building things.

It’s the year of the Linux Headset!

Anthropic will sign BAAs. They are HIPAA compliant (we used them.)

I understand people’s hesitation but the business agreements are different. The business risks of misusing HIPAA data is not only being banned from a massive enterprise market (last I checked, there were about 1.2m jobs related to claims billing and adjudication) and significant legal repercussions.

Other companies like AWS also handle HIPAA data. Are we afraid they’re stealing it? I don’t believe it, nor that Anthropic is training on HIPAA data.

OpenAI… I will never trust them.